Volumetric error modeling and accuracy improvement by parameter identification of a compound machine tool

抄録

This paper presents a systematic method for kinematic modeling and improving the positioning accuracy of a compound machine tool. With the configuration of model frames and the adjustment of the link offset parameters, the position and orientation of the tool center point (TCP) are measured conveniently by a laser tracker, and the forward kinematic solution of the machine model is provided. Through the Levenberg-Marquardt (L-M) algorithm combined with chi-square fitting, the maximum likelihood estimators of the model parameters are obtained. The identified parameters indicate a certain squareness error between the linear and adjacent rotary axes, and each coordinate axis has a certain angular error. The link length parameter also has a slight error. The calibration result shows that the average position and orientation error of the machine tool are 0.03999 mm and 6.571 × 10^-4 rad respectively, which are 79.6% and 44.9% lower than the initial error, indicating that the volumetric accuracy of the machine tool has been greatly improved through parameter identification. Compared with rigid body kinematics or screw theory modeling, the Denavit-Hartenberg (D-H) combined with Hayati-Mirmirani (H-M) modeling method used by the compound machine tool has a clear geometric meaning of the model parameters, and there is no need to measure the single geometric error of the coordinate axis. It has the advantages of fewer modeling steps and short test time. The volumetric error modeling, accuracy measurement, and parameter identification proposed in this paper are beneficial in improving the volumetric accuracy of machine tools with special structure.